Skip to main content

Warehouse Efficiency

The Warehouse Efficiency page at /finops/warehouses is the SQL warehouse-specific view of compute spend and utilisation. Use it to spot warehouses that are over-provisioned, idle too often, or running 24/7 without auto-stop.

Per-warehouse cards​

Each warehouse gets a card with:

  • Name + size — Small, Medium, Large, 2X-Large, …
  • Type — Pro / Serverless / Classic
  • State — running / stopped / starting
  • Auto-stop — minutes
  • Cost (window) — total over 30d
  • Avg utilisation %
  • Idle hours per day — when running but no queries

Sort and filter​

Sort by cost descending to find expensive warehouses. Filters:

  • Size
  • Type (Pro / Serverless)
  • State
  • Owner team
  • Min cost threshold

Utilisation chart​

Click any warehouse to expand a chart showing query concurrency over the last 7 days. Look for:

  • Tall sustained peaks → consider scaling up the warehouse
  • Many tall but brief peaks → already at the right size
  • Consistent low load → consider scaling down
  • Long idle gaps with running state → enable auto-stop or shorten it

Recommendations​

Per-warehouse hints (also surface on Recommendations):

  • Reduce auto-stop from 30m to 10m — saves $X/month based on observed idle pattern
  • Right-size from Large to Medium — avg utilisation is 24%
  • Switch to Serverless — bursty workload is a fit, est. savings $Y
  • Switch to Pro from Classic — Pro adds Photon at no extra cost for many query types

Bulk actions​

Multi-select warehouses for:

  • Set auto-stop — apply same value across selection
  • Resize — apply same target size
  • Stop — emergency stop
GET /finops/warehouses?days=30
POST /finops/warehouses/bulk
{
  "warehouse_ids": [...],
  "action": "set_auto_stop",
  "params": { "minutes": 10 }
}

API​

GET /finops/warehouses
GET /finops/warehouses/{id}
GET /finops/warehouses/{id}/utilisation?days=7
POST /finops/warehouses/{id}/right-size